© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 1517
    Prev
    Next
    INDEX
    Explanations

    references to a specific year, specifically \\"year 9\\"

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    651
    +0.14
    0.6%
    146
    +0.10
    0.5%
    436
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    651
    +0.14
    0.08
    306
    +0.10
    0.06
    436
    +0.10
    0.05
    Negative Logits
     textual
    -0.75
     cumbers
    -0.74
     zipper
    -0.72
     chains
    -0.64
    achus
    -0.63
    ogle
    -0.62
     typew
    -0.61
     encl
    -0.59
     fences
    -0.58
     sympath
    -0.58
    POSITIVE LOGITS
    long
    1.09
    nings
    1.00
    book
    0.94
    ning
    0.87
    ns
    0.84
     olds
    0.84
     anniversary
    0.83
     ago
    0.78
    ned
    0.78
    books
    0.78
    Activations Density 0.170%

    No Known Activations